Unemployment in the teens and trajectories of alcohol consumption in adulthood
Abstract
Objectives The unemployed are assumed to adopt unhealthy behaviours, including harmful use of alcohol. This study sought to elucidate the relations between unemployment before age 21 years and consumption of alcohol from 21 to 42 years. The design was based on the conception of youth as a sensitive period for obtaining ‘drinking scars’ that are visible up to middle age. Setting The Northern Swedish Cohort Study has followed up a population sample from 1981 to 2007 with five surveys. Participants All pupils (n=1083) attending the last year of compulsory school in Luleå participated in the baseline survey in classrooms, and 1010 of them (522 men and 488 women) participated in the last follow-up survey that was conducted at classmate reunions or by post or by phone. Outcome measure The trajectory of alcohol consumption from 21 to 43 years, obtained with latent class growth analyses, was scaled. Results Men were assigned to five and women to three consumption trajectories. The trajectory membership was regressed on accumulation of unemployment from 16 to 21 years, with multinomial logistic regression analyses. The trajectory of moderate consumption was preceded by lowest exposure to unemployment in men and in women. With reference to this, the relative risk ratios for high-level trajectory groups were 3.49 (1.25 to 9.79) in men and 1.41 (0.74 to 2.72) in women, but also the trajectories of low-level consumption were more probable (relative risk ratio 3.18 (1.12 to 9.02) in men and 2.41 (1.24 to 4.67) in women). Conclusions High-level alcohol consumption throughout adulthood is, particularly among men, partly due to ‘scars’ from youth unemployment, particularly in men, but there are also groups of men and women where unemployment in the teens predicts a trajectory of low consumption.
Full text
Unemploymen in he eens
and ajec o ies o alcohol consump ion
in adul hood
Pekka Vi anen,
1,2
Tomi Lin onen,
3
Hugo Wes e lund,
4
Tapio Nummi,
5
U ban Janle ,
2
Anne Hamma s öm
2
To ci e: Vi anen P,
Lin onen T, Wes e lund H,
e al. Unemploymen in he
eens and ajec o ies o
alcohol consump ion in
adul hood. BMJ Open
2016;6:e006430.
doi:10.1136/bmjopen-2014-
006430
▸P epublica ion his o y o
his pape is a ailable online.
To iew hese iles please
isi he jou nal online
(h p://dx.doi.o g/10.1136/
bmjopen-2014-006430).
Recei ed 22 June 2015
Re ised 4 Sep embe 2015
Accep ed 21 Oc obe 2015
1
Uni e si y o Tampe e,
Ins i u e o Ad anced Social
Resea ch and School o
Heal h Sciences, Tampe e,
Finland
2
Depa men o Public Heal h
and Clinical Medicine, Umeå
Uni e si y, Umeå, Sweden
3
The Finnish Founda ion o
Alcohol S udies, Helsinki,
Finland
4
Uni e si y o S ockholm,
S ess Resea ch Cen e,
S ockholm, Sweden
5
Uni e si y o Tampe e,
School o Heal h Sciences,
Tampe e, Finland
Co espondence o
D Pekka Vi anen;
[email p o ec ed]i
ABSTRACT
Objec i es: The unemployed a e assumed o adop
unheal hy beha iou s, including ha m ul use o
alcohol. This s udy sough o elucida e he ela ions
be ween unemploymen be o e age 21 yea s and
consump ion o alcohol om 21 o 42 yea s. The
design was based on he concep ion o you h as a
sensi i e pe iod o ob aining ‘d inking sca s’ ha a e
isible up o middle age.
Se ing: The No he n Swedish Coho S udy has
ollowed up a popula ion sample om 1981 o 2007
wi h i e su eys.
Pa icipan s: All pupils (n=1083) a ending he las
yea o compulso y school in Luleå pa icipa ed in he
baseline su ey in class ooms, and 1010 o hem (522
men and 488 women) pa icipa ed in he las ollow-up
su ey ha was conduc ed a classma e eunions o by
pos o by phone.
Ou come measu e: The ajec o y o alcohol
consump ion om 21 o 43 yea s, ob ained wi h la en
class g ow h analyses, was scaled.
Resul s: Men we e assigned o i e and women o
h ee consump ion ajec o ies. The ajec o y
membe ship was eg essed on accumula ion o
unemploymen om 16 o 21 yea s, wi h mul inomial
logis ic eg ession analyses. The ajec o y o mode a e
consump ion was p eceded by lowes exposu e o
unemploymen in men and in women. Wi h e e ence
o his, he ela i e isk a ios o high-le el ajec o y
g oups we e 3.49 (1.25 o 9.79) in men and 1.41
(0.74 o 2.72) in women, bu also he ajec o ies o
low-le el consump ion we e mo e p obable ( ela i e
isk a io 3.18 (1.12 o 9.02) in men and 2.41 (1.24 o
4.67) in women).
Conclusions: High-le el alcohol consump ion
h oughou adul hood is, pa icula ly among men,
pa ly due o ‘sca s’ om you h unemploymen ,
pa icula ly in men, bu he e a e also g oups o men
and women whe e unemploymen in he eens p edic s
a ajec o y o low consump ion.
BACKGROUND
A lo o esea ch has been de o ed o show
ha unemploymen inc eases in ake o
alcohol, binge d inking and diagnosed
alcohol use diso de s.
1–4
The esea ch is
based on he heo e ical easoning ha
losing one’s job and being jobless may be
accompanied by condi ions— o example,
inc eased dis ess, wo sening men al heal h,
eelings o us a ion and shame, loss o
s uc u es o daily li e and ensions in amily
and o he social ela ions— ha p edispose
indi iduals o inc ease hei d inking.
2
The e
a e findings suppo ing his logic, bu he
e idence o unemploymen as a cause o
d inking seems o be less consis en han he
e idence o d inking as a cause o unemploy-
men .
3
The e o e, i is easonable o assume
ha he e a e also condi ions ha may
dec ease d inking: in addi ion o educed
income, unemploymen means, in any case,
lack o job- ela ed s ess and also, in some
cases, dis ance om he ‘we ’cul u e ep e-
sen ed by wo kma es. Indeed, wo king seems
o be associa ed wi h high consump ion, a
leas in adolescence.
5
Unemploymen is qui e common a ound
he en y o wo k li e.
67
Pa ly, pe iods o
you h unemploymen may be conside ed as
ansien and ela i ely unimpo an ic ion
in he a achmen o he labou ma ke . On
he o he hand, unemploymen in he eens
S eng hs and limi a ions o his s udy
▪This was a long (27 yea s) ollow-up o a popu-
la ion coho wi h i ually o al pa icipa ion in
he su eys.
▪Owing o he high pa icipa ion a e, he e was
no d inking- ela ed sample a i ion.
▪By i ue o b eaking he a e age de elopmen
down o ajec o ies, he s udy cla i ied and
quan i ied he dual de elopmen : exposu e o
you h unemploymen inc eases he p obabili y o
bo h high and low-le el consump ion o alcohol
in adul hood.
▪The indings may be limi ed o Sweden and
simila socie ies.
Vi anen P, e al.BMJ Open 2016;6:e006430. doi:10.1136/bmjopen-2014-006430 1
Open Access Resea ch
g oup.bmj.com on May 18, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om
may be dele e ious, because i occu s du ing a phase o
he li e cou se ha is impo an o he de elopmen o
independence and iden i y o an indi idual. In addi ion
o—and also due o—causing li elong ‘wage penal ies’
8
and pe manen ‘sca s’in he labou ma ke a achmen
and employabili y,
9
unemploymen du ing his sensi i e
pe iod may lea e ‘sca s’, in e ms o socialisa ion, and a
li es yle o isky heal h beha iou s, as well as in e ms o
lowe ed men al well-being and may ul ima ely ha m
soma ic heal h. In his con ex , he ela ion o you h
unemploymen wi h alcohol consump ion in adul hood
poses an essen ial ques ion.
A s udy
10
ha ollowed up 17–19-yea -old No wegians
o 3 yea s ound ha exposu e o unemploymen was
no associa ed wi h alcohol consump ion (measu ed as
cl o pu e alcohol in 1 yea ) when con olling o he
in ake a baseline. In con as , a s udy om he USA
showed ha being hea y d inke (consuming mo e han
fi e d inks on one occasion) a he age o 27–35 yea s
was associa ed wi h exposu e o unemploymen du ing
he p eceding 12-yea pe iod
11
Acco ding o a B i ish
s udy among men,
12
he associa ion o high-alcohol con-
sump ion a he age o 33 yea s (belonging o he op
fi h) wi h unemploymen expe ienced be ween he ages
16 o 33 yea s was non-significan , whe eas unemploy-
men du ing he p eceding yea was associa ed wi h
hea y d inking.
The p esen s udy is based on a 26-yea ollow-up o
he No he n Swedish Coho .
13
An ea ly s udy wi h he
coho
14
ound ha , bo h among men and women,
hose exposed o unemploymen o mo e han
20 weeks a age 16–21 yea s consumed mo e alcohol a
21 yea s, compa ed wi h hose ha ing less o no expos-
u e. The associa ions be ween exposu e and consump-
ion emained significan a e adjus men s o
consump ion a 16 yea s and a se o socioeconomic
ac o s. In he s udy based on ollow-up un il he age o
30 yea s,
15
he pa icipan s we e classified in oan ea ly
unemploymen g oup (unemployed o e 6 mon hs
be ween he ages o 16 and 21 yea s, i espec i e o la e
unemploymen ), he la e unemploymen g oup
(unemployed o e 18 mon hs be ween he ages o 22
and 30 yea s and unemployed less han 6 mon hs
be ween he ages o 16 and 21 yea s) and he e e ence
g oup consis ing o hose who did no ulfil hese c i-
e ia. A dicho omous measu e (spli a 75 h cen ile, sep-
a a ely o men and women) was used as indica o o
high consump ion o alcohol e sus none. Th oughou
he ollow-up, he figu es o high consump ion we e
g ea es in he ea ly unemploymen g oups. A e adjus -
ing o consump ion a baseline, howe e , nei he in he
ea ly no la e unemploymen g oup did he p obabili y
o being a high consume a he age o 30 yea s di e
significan ly om he e e ence g oup.
In his s udy, we we e asking how exposu e o
unemploymen in he 1980s, when pa icipan s o he
No he n Swedish Coho we e in hei eens, is
eflec ed in hei consump ion o alcohol in he 1990s
and in he 2000s, up o he age o 42 yea s. Ou hypo h-
esis was ha , in spi e o he non-significan findings o
he p e ious s udy, an analysis based on longe - e m
ollow-up and iden ifica ion o longi udinal pa e ns o
consump ion would e eal ha you h unemploymen
p edic s high consump ion in adul hood bo h among
men and women.
DATA AND VARIABLES
The No he n Swedish Coho S udy was se up in 1981
wi h a ques ionnai e o all pupils (n=1083) who
a ended he las yea o compulso y school (age
16 yea s) in all nine schools in a middle-sized municipal-
i y in No he n Sweden. Follow-up su eys we e con-
duc ed in 1983 (age 18 yea s), 1987 (age 21 yea s), 1995
(age 30 yea s) and 2007 (age 42 yea s). The esponse
a es we e e y high: a he 26-yea ollow-up in 2007, a
o al o 522 men and 488 women (n=1010, 94% o hose
s ill ali e) con inued o pa icipa e.
The baseline su ey ook place in class ooms. The
ques ionnai e con ained que ies on soma ic and men al
heal h, amily backg ound, school pe o mance and
en i onmen and, as pa o heal h- ela ed beha iou s,
he use o alcohol. The pupils we e asked o a e hei
consump ion o alcohol as equency o d inking occa-
sions and a e age in ake o bee (numbe o bo les),
wine (numbe o glasses) and s ong be e ages (numbe
o d inks) on each occasion. This ques ion on alcohol
use was included in subsequen su eys ha ook place
a classma e eunions, and hose who did no a end
we e su eyed by pos o by elephone.
The ques ionnai e a 18 yea s o age included a ques-
ion abou o al du a ion o unemploymen in mon hs
du ing he p e ious 2 yea s, ha is, since he end o
compulso y school. The p incipal in es iga o (AH) con-
duc ed pe sonal in e iews abou leng h o unemploy-
men a he ollow-ups a 18 and 21 yea s wi h all hose
who had been unemployed du ing he pe iod. In he
ques ionnai e a 21 yea s, he e was a ma ix ha con-
sis ed o columns ep esen ing sp ing semes e , summe
and au umn semes e du ing he p e ious 3 yea s, and
ows ep esen ing di e en labou ma ke posi ions. I
he esponden ma ked only he posi ion ‘unemployed’,
he o she was classified as ha ing been unemployed o
he whole 6-mon h pe iod; i he e we e, in addi ion,
some o he posi ion(s), he numbe o unemploymen
mon hs was ob ained by di iding du a ion o he pe iod
by numbe o op ions. Values o he pe iods plus he
alues o ages 16 o 18 yea s we e hen summed in o
he a iable ‘mon hs o you h unemploymen ’.
The 1981 su ey included ques ions ega ding he
occupa ion o he mo he and he a he o he espond-
en . Applying he Swedish classifica ion sys em,
16
he
occupa ion o each pa en was classified in o social
g oup 1 (including en ep eneu s and highe whi e-
colla wo ke s), social g oup 2 (including lowe whi e
colla wo ke s) and social g oup 3 (including manual
2Vi anen P, e al.BMJ Open 2016;6:e006430. doi:10.1136/bmjopen-2014-006430
Open Access
g oup.bmj.com on May 18, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om
and non-manual wo ke s). The pa en al social class was
defined as ha ing no, one o wo pa en s om social
g oup 3. The ques ion abou soma ic and men al heal h
o he a he and mo he specified alcoholism as one o
he heal h p oblems; he eplies we e combined in o a
‘pa en al alcoholism’ a iable.
T uancy a 16 yea s o age was included in he analyses
as an indica o o ex e nalised beha iou . Replies o he
fi e-op ion ques ion abou absence om school wi hou
illness o a e asking o and ha ing been g an ed pe -
mission we e dicho omised as ne e /a ew imes in a
semes e .
Measu es o alcohol consump ion
Repo ed amoun s and equencies o in ake o be -
e ages we e calcula ed in o yea ly a e age in ake in cen-
ili es o absolu e alcohol a 21, 30 and 42 yea s, which
was hen used o cons uc wo di e en a iables o
alcohol consump ion. Fi s , hea y d inking a he 21, 30
and 42 yea s was defined as belonging o he uppe con-
sump ion qua ile, sepa a ely o men and women.
Second, he e ogenei y in he longi udinal de elopmen
o he consump ion om 21 o 42 yea s was analysed by
la en class g ow h analysis.
17 18
The consump ion ajec-
o ies we e iden ified using he Flexmix package
19
in R
s a is ical so wa e. Fi s deg ee polynomial (linea )
models we e fi ed and p io p obabili ies es ima ed
using he me hod o maximum likelihood. The Bayesian
In o ma ion C i e ia (BIC) sugges ed h ee ajec o ies
o women and fi e ajec o ies o men as p e e ed
solu ions.
Among women (figu e 1), 47% we e assigned o he
ajec o y cha ac e ised by s eady ‘mode a e le el’con-
sump ion h oughou he ollow-up. The g oup
membe s comp ising 31% o women we e ‘high le el’
consume s al eady a he baseline and sligh ly inc eased
consump ion by age, and 22% o women we e assigned
o he ‘low le el’ ajec o y.
Among men (figu e 2), abou hal we e assigned
ei he o he ‘high le el’(27%) o o he ‘inc easing o
high le el’(22%) ajec o y g oup, he o me showing
a sligh decline o consump ion, while among he la e
he consump ion a he end o he ollow-up was highes
o all. The o he hal o men was dis ibu ed e enly o
he ‘low le el’ ajec o y (18%), he ‘inc easing o mod-
e a e le el’ ajec o y (16%) and he ‘dec easing o mod-
e a e le el’ ajec o y (17%).
S a is ics
The dis ibu ion o alcohol consump ion was highly
skewed o he igh , and he e o e log ans o med
alues o he yea ly alcohol consump ion we e used in
analyses o he d inking ajec o ies, as well as in adjus -
ing o he baseline consump ion (age 16 yea s), and
he eg ession analyses be ween you h unemploymen
and adul d inking. Bina y logis ic eg essions we e used
in analysing he associa ions o you h unemploymen
wi h hea y d inking a di e en ages, and mul inomial
logis ic eg essions in analysing i s associa ions o he a-
jec o ies. In addi ion o baseline consump ion, he ana-
lyses we e adjus ed o uancy, and social class and
alcoholism o he pa en s. Since he iden ified ajec o -
ies a e based on pos e io p obabili ies, he eg ession
analyses we e weigh ed wi h he p obabili y used (ie,
highes one) when assigning indi iduals o he iden ified
ajec o ies.
RESULTS
Associa ions o you h unemploymen wi h adul hood hea y
d inking
High exposu e o you h unemploymen p edic ed hea y
d inking a 21 yea s, in men and in women ( able 1),
independen ly o being a hea y d inke a baseline. A
30 yea s, he associa ion was s a is ically non-significan
e en in he unadjus ed analysis among men. Among
women, he OR was non-significan a e adjus men s. By
42 yea s, he associa ions o you h unemploymen wi h
hea y d inking we e s ill non-significan , and in men he
OR had e en d opped below uni y ( o 0.91 in he
unadjus ed and 0.94 in he adjus ed model).
Associa ions o you h unemploymen wi h adul hood
d inking ajec o y
The ajec o y g oup wi h lowes a e age exposu e o
you h unemploymen was chosen as he e e ence in he
mul inomial logis ic eg essions. The analyses showed
ha women on he ‘Mode a e le el’and men on he
‘Inc easing o mode a e le el’ ajec o y we e he leas
exposed. The ables we e cons uc ed wi h hese g oups
in he le column, and he es o he columns we e
a anged in he o de o inc easing a e age exposu e o
you h unemploymen o he ajec o y g oup.
Figu e 1 Illus a ions o he ajec o ies o alcohol
consump ion om 21 o 42 yea s among women (n=488) o
he No he n Swedish Coho .
Vi anen P, e al.BMJ Open 2016;6:e006430. doi:10.1136/bmjopen-2014-006430 3
Open Access
g oup.bmj.com on May 18, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om
Women
The findings o women a e displayed in able 2. The
e e ence g oup wi h ‘Mode a e le el’consump ion a-
jec o y had on he a e age been exposed o 2.40 mon hs
o you h unemploymen . The exposu e was highes
(3.54 mon hs) in he g oup assigned o he ‘Low le el’
ajec o y. ORs showed he di e ence be ween he ‘High
le el’and he ‘Mode a e le el’ ajec o y g oups as non-
significan , while he di e ence be ween he ‘Low le el’
and he ‘Mode a e le el’g oups was significan .
Mo eo e , in case o he compa ison be ween low and
mode a e le el, he OR g ew when consump ion a base-
line was added o he model.
Men
Among men, he a e age exposu e o you h unemploy-
men anged om 5.43 mon hs in he ‘High le el’
g oup o 1.34 mon hs in he g oup wi h ‘Inc easing o
mode a e le el’consump ion ajec o y ( able 3).
The ajec o y o he g oups ‘Inc easing o high le el’
and ‘Dec easing o mode a e le el’did no di e signifi-
can ly om he e e ence g oup, whe eas he ORs
showed he ‘Low le el’and ‘High le el’ ajec o y g oups
as significan ly di e en . O e all, he ORs d opped a e
adjus men s bu he esul pa e n, as well as s a is ical
significance o he findings, emained unchanged.
DISCUSSION
Acco ding o ea lie s udies o he No he n Swedish
Coho , high exposu e o unemploymen is associa ed
wi h high-alcohol consump ion among you h,
14
bu
you h unemploymen does no ha e an independen
long- e m e ec on hea y consump ion a 30 yea s.
15
In
he p esen s udy, we se ou o analyse in mo e de ail
he po en ial sca ing e ec s o you h unemploymen on
alcohol consump ion du ing he li e cou se, by u ilising
he longe ollow-up un il 42 yea s and ajec o y ana-
lyses. As a kind o baseline analysis, we eplica ed he
Figu e 2 Illus a ions o he
ajec o ies o alcohol
consump ion om 21 o 42 yea s
among men (n=522) o he
No he n Swedish Coho .
Table 1 Associa ions (ORs and 95% CIs) be ween exposu e o unemploymen be ween he ages o 16 and 21 yea s and
hea y consump ion o alcohol (being in he uppe qua ile) a he ages o 21, 30 and 42 yea s, in men and women
Men Women
Age (yea s) Unadjus ed Adjus ed* Unadjus ed Adjus ed*
21 2.47 (1.69 o 3.60) 2.04 (1.36 o 3.07) 2.22 (1.36 o 3.62) 1.82 (1.07 o 3.10)
30 1.29 (0.91 o 1.83) 1.21 (0.83 o 1.78) 1.71 (1.05 o 2.79) 1.53 (0.91 o 2.58)
42 0.91 (0.61 o 1.37) 0.94 (0.61 o 1.44) 1.14 (0.69 o 1.89) 1.04 (0.60 o 1.81)
*Adjus ed o social class and alcoholism o he pa en (s) and o alcohol consump ion and uancy a baseline (age 16 yea s).
Bold ype ace ep esen s s a is ical signi icance (p<0.05).
4Vi anen P, e al.BMJ Open 2016;6:e006430. doi:10.1136/bmjopen-2014-006430
Open Access
g oup.bmj.com on May 18, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om
s udy o Hamma s öm and Janle
15
in o he ex ended
se ing, and ound ha he e ec o you h unemploy-
men on hea y d inking con inues a enua ing up o
middle age in bo h gende s. Nex , we b oke down he
popula ion coho in o g oups wi h di e en adul hood
consump ion ajec o ies and ela ed hem wi h expos-
u e o you h unemploymen . The esul s sugges ed ha ,
among men wi h high-le el consump ion h oughou he
pe iod om ea ly adul hood o ea ly middle age, pa o
he consump ion may be due o ‘sca s’ om you h
unemploymen . Among women assigned o he ajec o y
o high-le el consump ion, co esponding sca ing was
less e iden . These findings suppo ou hypo hesis, a
leas in men. Howe e , we also ound ‘in e se sca ing’:
bo h among men and among women he e we e g oups
o conside able size ha assumed low-le el consump ion
ajec o ies in adul hood, in spi e o ha ing a his o y o
ela i ely high exposu e o you h unemploymen .
The whole-coho analysis is in line wi h he esul s o
he B i ish s udy, which ound ha hea y d inking was
associa ed wi h ecen bu no wi h ea lie exposu e o
unemploymen .
12
Resul s o he US s udy on hea y
d inking as he ou come,
11
howe e , ends o con adic
ou findings by showing ha he e is an associa ion
be ween exposu e o in olun a y unemploymen du ing
eens and ea ly adul hood on he one hand, and alcohol
consump ion in ea ly middle age on he o he .
In he ajec o y analyses, we chose o use unca e-
go ised consump ion figu es as he ou come. In his
espec , he p esen s udy is simila o he ea lie men-
ioned s udy om No way,
10
which, howe e , used
a e age consump ion whe eas he ou come in ou s udy
was a spec um o consump ion ajec o ies. In he ligh
o ou esul s, he No wegian finding, acco ding o
which exposu e o unemploymen in he eens does no
p edic consump ion, can be unde s ood as a mix u e o
di e en ial associa ions. In o he wo ds, he e seems o
be a genuine dual associa ion be ween unemploymen and
use o alcohol: among hose exposed o you h unemploy-
men he e a e bo h hose wi h high and hose wi h low
consump ion in adul hood. The e o e, conclusions d awn
om analyses o a e age de elopmen may con ey an
o e ly uni o m and simplified depic ion o eali y.
Use o he amoun o alcohol— ega dless o he
pa e n o consump ion—as he ou come, is in line wi h
he ideology o ajec o y analysis, which is no o p efix
Table 2 Exposu e o you h unemploymen ( om 16 o 21 yea s) and alcohol consump ion ajec o ies om 21 o 42 yea s in
women
Consump ion ajec o y
Mode a e le el
n=224
High le el
n=150
Low le el
n=106
Mon hs o you h unemploymen (mean) 2.40 3.15 3.54
OR (95% CI)* e 1.49 (0.82 o 1.71) 1.87 (1.02 o 3.40)
OR (95% CI)† e 1.72 (0.93 o 2.20) 1.89 (1.01 o 3.55)
OR (95% CI)‡ e 1.41 (0.74 o 2.72) 2.41 (1.24 o 4.67)
ORs wi h 95% CIs o belonging o he ajec o y g oup in o de o inc easing exposu e, ha is, wi h ze o exposu e o he ‘mode a e le el’
g oup as he e e ence.
*Unadjus ed.
†Adjus ed o social class and alcoholism o he pa en (s).
‡Adjus ed o (2) and alcohol consump ion and uancy a baseline (age 16 yea s).
Bold ype ace ep esen s s a is ical signi icance (p<0.05).
Table 3 Exposu e o you h unemploymen ( om 16 o 21 yea s) and alcohol consump ion ajec o ies om 21 o 42 yea s o
age in men
Consump ion ajec o y
Inc easing o
mode a e le el
n=86
Inc easing o high
le el
n=133
Dec easing o
mode a e le el
n=95
Low le el
n=84
High le el
n=117
Mon hs o you h
unemploymen
(mean)
1.34 1.86 2.40 4.16 5.43
OR (95% CI)* e 1.56 (0.51 o 4.82) 2.21 (0.73 o 6.76) 3.76 (1.34 o 10.6) 4.85 (1.76 o 13.4)
OR (95% CI)† e 1.53 (0.49 o 4.77) 2.11 (0.68 o 6.50) 3.73 (1.31 o 10.6) 5.02 (1.80 o 14.0)
OR (95% CI)‡ e 1.24 (0.40 o 3.86) 1.66 (0.54 o 5.13) 3.18 (1.12 o 9.02) 3.49 (1.25 o 9.79)
ORs wi h 95% CIs o belonging o he ajec o y g oup in o de o inc easing exposu e, ha is, wi h ze o exposu e o he ‘inc easing o
mode a e le el’g oup as he e e ence.
*Unadjus ed.
†Adjus ed o social class and alcoholism o he pa en (s).
‡Adjus ed o (2) and alcohol consump ion and uancy a baseline (age 16 yea s).
Bold ype ace ep esen s s a is ical signi icance (p<0.05).
Vi anen P, e al.BMJ Open 2016;6:e006430. doi:10.1136/bmjopen-2014-006430 5
Open Access
g oup.bmj.com on May 18, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om
o ca ego ise he phenomenon o in e es bu , a he , o
see wha kind o ajec o ies he da a include and o
in e p e hem—in a sense, o ell associa ed ‘de elop-
men al s o ies’. Mo eo e , we a gue ha , by i ue o
using consump ion as he ou come a iable, he p esen
s udy e ealed phenomena ha ha e emain unde ec ed
in analyses wi h p oblema ic d inking ( equency o
occasions, binge d inking, e c) as he ou come. Despi e
he di e en numbe s o ajec o ies, we ound ha in
bo h gende s he ajec o y g oups associa ed wi h lowes
you h unemploymen ha e been cha ac e ised as d ink-
ing ‘mode a ely’. In women, his g oup includes almos
hal o he coho (44%), while in men he p opo ion
is smalle (16%).
We included consump ion and uancy a baseline in
he analyses when he coho was s ill a ending com-
pulso y school and had no ye been a isk o exposu e
o unemploymen bu was acing he en y o labou
ma ke —o o he ma ke o pos basic educa ion. A
his poin o ime in li e, uancy can be seen as an
indica o o ela i ely low in e es and mo i a ion o
pu sue u he s udies, as well as being a symp om o
ex e nalised beha iou . Some o he inhe i ed isk o
alcohol use diso de s was aken in o accoun wi h he
pa en al alcoholism a iable, and in line wi h he
design o he s udy, we conside ed i mo e app op ia e
o con ol o he socioeconomic condi ions by pa en al
social class han by own occupa ional posi ion a ained
du ing he ollow-up. Fo simila easons, we did no
include adul hood unemploymen in he mul i a ia e
models.
Ou s udy e ealed, on i s pa , a complexi y o he
associa ions be ween alcohol consump ion and wo k li e
cou se. Fo ins ance, he ajec o y g oup o men wi h
‘inc easing o high le el’consump ion had su e ed ela-
i ely li le om you h unemploymen . This aises he
ques ion o consump ion pa e ns: maybe hese indi i-
duals end o d ink a he equen ly, bu small amoun s
on each occasion, and he e o e can keep a s able
labou ma ke posi ion. Mo eo e , we can ques ion how
a he low-le el ajec o y can be seen as an indica o o
an ad an ageous li e cou se: maybe hese men and
women a e less socially ac i e and he e o e d ink less,
o hey canno a o d o d ink so much (especially o go
ou ), and some migh e en abs ain om d inking
because hey ha e had ea lie p oblems wi h alcohol
and belie e ha d inking could ha m hei chances in
he labou ma ke . The ajec o ies p o ide a no el s a -
ing poin o u u e s udies conce ning ‘disad an aged
low consume s’, bu also po en ial ‘ad an aged high
consume s’.
The long ollow-up ime and high pa icipa ion a e
a e e iden s eng hs o ou s udy. The sample also
included hose ma ginalised indi iduals wi h mo e
se e e alcohol abuse, and he e o e he ajec o ies ep-
esen a eal dis ibu ion o he popula ion.
Measu emen o unemploymen by a ques ionnai e,
ins ead o egis e da a, also can be conside ed a
s eng h, as sel - epo ed unemploymen indica es ha
he unemploymen expe ience has been significan . A
limi a ion is ha he findings may be conside ed locali y
specific, al hough i has been shown ha he coho in
many espec s is ep esen a i e o he whole coun y.
11
Wes e n coun ies do no di e in he main s uc u es o
hei educa ional sys ems: a e compulso y school, some
g adua es en e in o pos basic educa ion while o he s
en e he labou ma ke . Wi h espec o his, ou esul s
can be gene alised, al hough in Sweden, en y in o edu-
ca ion may be mo e common han in mos o he
coun ies.
The p esen s udy is pa o he esea ch p ojec
‘Unemploymen Sca ing’. Acco ding o he sca ing
hypo hesis, he beha iou al changes ha ake place
du ing unemploymen do no change ins an aneously a
he end o unemploymen , bu es o a ion akes ime
and may emain incomple e o e a long pe iod o ime.
The na u e o sca s may be assumed o depend, in add-
i ion o he amoun o exposu e o unemploymen , on
he momen o exposu e du ing he li e cou se. The
yea s o en y in o he labou ma ke may be a pe iod o
pa icula sensi i i y o he consequences o unemploy-
men
20
as well as o ob aining sca s o unemploymen .
The sca ing app oach was also he eason we did no
analyse po en ial di e ences in he s eng h o he asso-
cia ions be ween d inking and ac ual unemploymen a
di e en ages. This in e es ing ques ion emains a opic
o u u e esea ch.
Wi h espec o policy implica ions, measu es ha e-
duce you h unemploymen migh educe he ela i ely
la ge p opo ion (abou hal o he s udied coho )
assuming high-le el consump ion in adul hood. Al hough
he p opo ion o hose assuming low-le el d inking a-
jec o ies may also dec ease, he ne e ec would p obably
be educ ion o consump ion and also o p oblema ic use
o alcohol.
Con ibu o s PV concei ed he s udy, pa icipa ed in i s design and d a ed
he manusc ip . TL and HW pa icipa ed in design o he s udy and
in e p e a ion o he esul s, and commen ed on he manusc ip . TN
pa icipa ed in he design o he s udy, pe o med he ajec o y analyses
and d a ed co esponding pa s o he manusc ip . UJ and AH concei ed
he s udy, pa icipa ed in i s design and coo dina ion, and con ibu ed o
d a ing o he manusc ip . All he au ho s ead and app o ed he inal
manusc ip .
Funding The s udy has been inanced by The Swedish Resea ch Council o
En i onmen , Ag icul u al Sciences and Spa ial Planning dn 259-2012-37 and
by he Academy o Finland (g an numbe 132668).
Compe ing in e es s None decla ed.
E hics app o al Regional E hical Re iew Boa d in Umeå, Sweden.
P o enance and pee e iew No commissioned; ex e nally pee e iewed.
Da a sha ing s a emen No addi ional da a a e a ailable.
Open Access This is an Open Access a icle dis ibu ed in acco dance wi h
he C ea i e Commons A ibu ion Non Comme cial (CC BY-NC 4.0) license,
which pe mi s o he s o dis ibu e, emix, adap , build upon his wo k non-
comme cially, and license hei de i a i e wo ks on di e en e ms, p o ided
he o iginal wo k is p ope ly ci ed and he use is non-comme cial. See: h p://
c ea i ecommons.o g/licenses/by-nc/4.0/
6Vi anen P, e al.BMJ Open 2016;6:e006430. doi:10.1136/bmjopen-2014-006430
Open Access
g oup.bmj.com on May 18, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om
REFERENCES
1. Fo cie M. Unemploymen and alcohol abuse: a e iew. J Occup
En i on Med 1988;30:246–51.
2. Henkel D. Unemploymen and subs ance use: a e iew o he
li e a u e (1990–2010). Cu D ug Abuse Re 2011;4:4–27.
3. de Goeij MC, Suh cke M, To olu i V, e al. How economic c ises
a ec alcohol consump ion and alcohol- ela ed heal h p oblems: a
ealis sys ema ic e iew. Soc Sci Med 2015;131:131–46.
4. Paljä i T, Ma ikainen P, Pensola T, e al. Li e cou se ajec o ies o
labou ma ke pa icipa ion among young adul s who expe ienced
se e e alcohol- ela ed heal h ou comes: a e ospec i e coho s udy.
PLoS ONE 2015;10:e0126215.
5. Kou onen A, Lin onen T. Adolescen pa - ime wo k and hea y
d inking in Finland. Addic ion 2002;97:311–18.
6. Eu o ound. Mapping you h ansi ions in Eu ope. Luxembou g:
Publica ions O ice o he Eu opean Union 2014. h p://eu o ound.
eu opa.eu/publica ions/ epo /2014/labou -ma ke /
mapping-you h- ansi ions-in-eu ope
7. Eu o ound. Young people and empo a y employmen in Eu ope.
Luxembou g: Publica ions O ice o he Eu opean Union 2014. h p://
eu o ound.eu opa.eu/obse a o ies/emcc/compa a i e-in o ma ion/
young-people-and- empo a y-employmen -in-eu ope
8. A ulampalam W. Is unemploymen eally sca ing?
E ec s o unemploymen expe iences on wages. Econ J
2001;111:585–606.
9. Eliasson M, S o ie D. La en o las ing sca s? Swedish e idence on
he long- e m e ec s o job displacemen . J Labo Econ
2006;24:831–56.
10. Hamme T. Unemploymen and use o d ug and alcohol among
young people: a longi udinal s udy in he gene al popula ion. B J
Addic 1992;87:1571–81.
11. Mossakowski KN. Is he du a ion o po e y and unemploymen a
isk ac o o hea y d inking? Soc Sci Med 2008;67:947–55.
12. Mon gome y S, Cook D, Ba ley M, e al. Subs ance abuse.
Unemploymen , ciga e e smoking, alcohol consump ion and body
weigh in young B i ish men. Eu J Public Heal h 1998;8:21–7.
13. Hamma s öm A, Janle U. Coho p o ile: he No he n Swedish
Coho . In J Epidemiol 2012;41:1545–52.
14. Janle U, Hamma s öm A. Alcohol consump ion among
unemployed you hs: esul s om a p ospec i e s udy. B J Addic
1992;87:703–14.
15. Hamma s öm A, Janle U. Ea ly unemploymen can con ibu e o
adul heal h p oblems: esul s om a longi udinal s udy o school
lea e s. J Epidemiol Communi y Heal h 2002;56:624–30.
16. S a is ics Sweden. Swedish socio-economic classi ica ion (in
Swedish). S ockholm: S a is ics Sweden, 1983.
17. Nagin D. Analyzing de elopmen al ajec o ies: a semipa ame ic,
g oup-based app oach. Psychol Me hods 1999;4:139–57.
18. Nagin D. G oup-based modelling o de elopmen . London: Ha a d
Uni e si y P ess, 2005.
19. Leisch F. FlexMix: a gene al amewo k o ini e mix u e models and
la en class eg ession in R. J S a So w 2004;11:1–18.
20. Reine I, No o M, Hamma s öm A. Does he associa ion be ween ill
heal h and unemploymen di e be ween young adul s and adul s?
Resul s om a 14-yea ollow-up s udy wi h a ocus on psychological
heal h and smoking. Public Heal h 2004;118:337–45.
Vi anen P, e al.BMJ Open 2016;6:e006430. doi:10.1136/bmjopen-2014-006430 7
Open Access
g oup.bmj.com on May 18, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om
o alcohol consump ion in adul hood
Unemploymen in he eens and ajec o ies
Janle and Anne Hamma s öm
Pekka Vi anen, Tomi Lin onen, Hugo Wes e lund, Tapio Nummi, U ban
doi: 10.1136/bmjopen-2014-006430
2016 6: BMJ Open
h p://bmjopen.bmj.com/con en /6/3/e006430
Upda ed in o ma ion and se ices can be ound a :
These include:
Re e ences #BIBLh p://bmjopen.bmj.com/con en /6/3/e006430
This a icle ci es 16 a icles, 3 o which you can access o ee a :
Open Access
h p://c ea i ecommons.o g/licenses/by-nc/4.0/non-comme cial. See:
p o ided he o iginal wo k is p ope ly ci ed and he use is
non-comme cially, and license hei de i a i e wo ks on di e en e ms,
pe mi s o he s o dis ibu e, emix, adap , build upon his wo k
Commons A ibu ion Non Comme cial (CC BY-NC 4.0) license, which
This is an Open Access a icle dis ibu ed in acco dance wi h he C ea i e
se ice
Email ale ing box a he op igh co ne o he online a icle.
Recei e ee email ale s when new a icles ci e his a icle. Sign up in he
Collec ions
Topic A icles on simila opics can be ound in he ollowing collec ions
(1531)Public heal h (1510)Epidemiology
No es
h p://g oup.bmj.com/g oup/ igh s-licensing/pe missions
To eques pe missions go o:
h p://jou nals.bmj.com/cgi/ ep in o m
To o de ep in s go o:
h p://g oup.bmj.com/subsc ibe/
To subsc ibe o BMJ go o:
g oup.bmj.com on May 18, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om